JP2000105675A - Disk array device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent illegal data from being written into a disk array device by a fault of a bus connecting a disk array controller with respective disk devices. SOLUTION: Before writing write data (update data and check code thereof) transferred from a disk array controller 5 into a disk part 11, a disk control part 12 of the disk device checks the propriety of data through a write data check means 1212. The write data check means 1212 is provided with a check code generator 12121 for generating the check code from update data in the write data and a comparator 12122 for comparing the generated check code with the check code in the write data. Only when both the check codes are coincident, writing to the disk part 11 is performed and when these check codes are not coincident, abnormal end is reported to the disk array controller 5.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk array device, and more particularly, to a disk array device having a function of checking the validity of write data.

[0002]

2. Description of the Related Art Generally, a disk array device comprises a plurality of disk devices and a disk array control device for controlling the plurality of disk devices. Each disk device includes a disk unit as a medium for storing data, and a disk control unit that writes and reads data to and from the disk unit in accordance with an instruction from the disk array control device. Of the plurality of disk devices, for example, one of them is used for parity and the rest is used for data, and even if one of the data disk devices fails, data can be recovered with the remaining disk devices. It has become. Each disk unit stores data and a check code for the data (for example, CRC (Cyclic Redundancy) in physical sector units.
Check) code), the disk control unit verifies the validity of the data with the check code stored in the physical sector when reading a certain physical sector, and if an error is found, Correct the data.

For example, in a disk array device having four disk devices 1, 2, 3, and 4 as shown in FIG. 5, for example, three disk devices 1, 2, and 3 are used for data and one disk device is used. The disk device 4 is used for parity.
Disk unit 1 in data disk units 1, 2, 3
Each of the physical sectors 1, 21, 31 stores data and a check code for the data. For example, as shown, data D0 and its check code CRC0 are stored in physical sector 0 of the disk unit 11 in the disk unit 21.
Physical sector 0 has data D1 and its check code C
RC1 stores data D in physical sector 0 of the disk unit 31.
2 and its check code CRC2 are stored, respectively. The disk unit 4 of the parity disk unit 4
1 stores parity data of data stored in the same physical sector of the data disk devices 1, 2, and 3, and a check code of the parity data. For example, the physical sector 0 of the disk unit 41 of the disk device 4 stores the parity data CD0 of the data D0, D1, and D2 stored in the physical sector 0 of the disk devices 1, 2, and 3.
12 and a check code CRC012 of the parity data CD012.

A set of data stored in the same physical sector of the data disk devices 1, 2, 3 is called logical data. For example, in the case of physical sector 0, data D0, D
The combination of 1 and D2 is one logical data. 1
When reading the entire logical data, the data of the physical sector 0 is simultaneously read in all the disk devices 1, 2, 3, and 4, and the data checked by the check code is used in each of the disk devices 1, 2, 3, and 3. , 4 to the disk array controller 5, where the disk array controller 5 checks with parity data. For example, when reading the entire logical data corresponding to physical sector 0, the validity of data D0 is verified from data D0 read from physical sector 0 and its check code CRC0 in disk device 1. Similarly, data D1, D2, and CD012 read from physical sector 0 in disk devices 2, 3, and 4, and check code CRC thereof.
, CRC2, and CRC012 from the data D1, D2,
The validity of CD012 is verified. Data D0, D1,
If an error is detected in D2 and CD012, data D0 and D are stored in each of the disk devices 1, 2, 3 and 4, if possible.
1, D2 and CD012 are corrected and sent to the disk array controller 5. In the disk array control device 5, the read data D0, D1, D2 and the parity data C
After verifying the validity of the logical data from D012, the logical data is sent to the host processor.

When reading only a part of the logical data, for example, the data of the physical sector 0 of the disk device 1, the disk array control device 5 operates only the disk device 1. Also at this time, the disk device 1 verifies the validity of the read data D0 using the check code CRC0.

On the other hand, at the time of updating data, the disk array control device 5 operates the disk device 1 for the corresponding data.
Data is written to the disk devices 2 and 3 and the parity disk device 4. For example, when updating all the logical data corresponding to the physical sector 0 to the data D0 ', D1', D2 ', the disk array controller 5 sets the data D0'
, The check code CRC1 'from the data D1', the check code CRC2 'from the data D2', and the parity data CD01 from the data D0 ', D1', D2 '.
2 ′ and a check code CRC012 ′ of the parity data CD012 ′. Then, data D0 'and check code CRC0' are stored in physical sector 0 of disk device 1, data D1 'and check code CRC1' are stored in physical sector 0 of disk device 2, and data D2 'are stored in physical sector 0 of disk device 3. Check code C
RC2 'is stored in the physical sector 0 of the disk device 4 with the parity data CD012' and the check code CRC012 '.
Are written to each of the disk devices 1, 2, 2,
A write request is sent to 3, 4.

When updating a part of the logical data, the disk array controller 5 controls the data recorded on the corresponding disk devices 1, 2, 3 and the parity disk device 4. Read and update,
Perform rewriting. For example, when updating only the data D0 of the physical sector 0 of the disk device 1 to D0 ', the disk array control device 5 sends the data D0
And parity data C from the disk device 4
D012 is read, and from these and new data D0 ',
The parity data CD012 'is generated and the check code CRC01 of the parity data CD012' is generated.
Generate 2 '. Further, the check code CRC0 'is generated from the new data D0'. And new data D
A write request is sent to each of the disk devices 1 and 4 to write 0 'and the check code CRC0' to the physical sector 0 of the disk device 1 and write the parity data CD012 'and the check code CRC012' to the physical sector 0 of the disk device 4. I do.

[0008]

As described above, the conventional disk array device stores data and its check code in the physical sector unit of each disk device.
Although the reliability of the data is further improved, the verification of the data by the check code is limited at the time of reading, and is not performed at the time of writing the data. Therefore, the disk array control device 5 and each of the disk devices 1, 2, 2,
When a failure occurs in the bus connecting the buses 3 and 4, there is a problem that incorrect data is written to the disk device. For example, when the data D0 of the physical sector 0 of the disk device 1 is updated to the data D0 ′ as described above, the disk array control device 5 sends the new data D0 ′ and its check code CRC0 ′ to the disk device 1. 1 is transmitted through a bus between the two, but when a failure occurs in the bus and, for example, the new data D0 'is garbled into D0 ", the data D0" and the check code CRC0' are stored in the physical sector 0. Is done. Since such illegal data is not checked until it is read out, especially when a large amount of data is updated continuously, illegal data continues to be written to many parts of the disk device, and in the worst case, recovery becomes impossible. I will.

On the other hand, as a data check method at the time of writing data in a disk array device, Japanese Patent Laid-Open No.
There is a technique described in Japanese Patent No. 66304. According to this technique, when four-path data sent from a host processor is converted into eight-path data in a disk array controller and divided into eight disk devices and written, the disk array controller converts the data. The original four-pass data is re-synthesized from the eight-pass data to independently generate each check data, and the generated check data and the check data accompanying the four-pass data sent from the host processor are generated. Are compared to check the data written to the disk device. According to this technique, it is possible to detect garbled data when a failure occurs in the data conversion unit inside the disk array control device. However, when there is a failure in the bus connecting the disk array control device and each disk device, data corruption cannot be detected, and the above-mentioned problem cannot be solved.

SUMMARY OF THE INVENTION An object of the present invention is to provide a disk array device capable of preventing unauthorized data from being written to a disk device due to a failure of a bus connecting the disk array control device and each disk device. To provide.

[0011]

A disk array device according to the present invention comprises a plurality of disk devices each comprising a disk control unit and a disk unit, and a disk array control device for controlling the plurality of disk devices. In the disk array device, the disk control unit of each disk device includes write data checking means for checking the validity of the data before writing the data transferred from the disk array control device to the disk unit. It is characterized by. More specifically, the write data check unit generates a check code from data to be written to a disk unit, and the check code generated by the unit and the check code attached to the data transferred from the disk array controller. Means for comparing with a check code.

In the present invention, when the validity of the data is confirmed by the write data check means, the disk control unit writes the data to the disk unit and then reports a normal end to the disk array control device. When the data abnormality is detected by the write data check means, an abnormal end is reported to the disk array controller without writing the data to the disk unit.

Further, the disk array control device according to the present invention has a configuration in which a retry process is performed when an abnormal end is reported from any disk device that has requested data writing.

In the thus constructed disk array device of the present invention, the disk control unit of the disk device to which the data and the check code have been sent from the disk array control device transmits the data to the disk unit. Before writing, the validity of the data is checked by comparing the check code uniquely generated from the data with the check code attached to the data transferred from the disk array controller by the write data checking means. Then, when the validity of the data is confirmed, writing of the data to the disk unit is executed, and then a normal end is reported to the disk array controller. When an abnormality of the data is detected, the writing of the data to the disk unit is performed. Report abnormal termination to the disk array controller without performing When an abnormal end is reported from any of the disk devices that have requested data writing, the disk array control device performs a retry process.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of a disk array device according to the present invention. The disk array device of the present embodiment includes a plurality of disk devices 1, 2, 3, 4 and a host processor 6
And a disk array control device 5 which controls the data transfer between the plurality of disk devices 1, 2, 3, and 4 and the host processor 6 in accordance with the contents of the input / output command from the host processor 6. The disk array controller 5 and the host processor 6 are connected by a host bus 7, and the disk array controller 5 and the individual disk devices 1, 2, 3, and 4 are connected to the disk buses 8-1, 8-2, and 8-3. , 8-4.

Each of the disk units 1, 2, 3, and 4 has a disk unit 11, 21, 3, which is a medium for storing data.
1 and 41 and disk controllers 12 and 2 for writing and reading data to and from the disk units 11, 21, 31 and 41 in accordance with instructions from the disk array controller 5.
2, 32, and 42. In the case of the present embodiment, the plurality of disk devices 1, 2, 3, 4 are such that the disk device 4 is used for parity and the remaining disk devices 1, 2, 3, 3 are used.
Are used for data, and even if one of the disk devices fails, the data can be recovered by the remaining disk devices. Also, each of the disk units 11, 21, 3
Reference numerals 1 and 41 denote data and a check code for the data (for example, CRC (Cyclic Re
(Dundancy Check) code).

For example, data D0 and its check code CRC0 are stored in physical sector 0 of disk unit 11, data D1 and its check code CRC1 are stored in physical sector 0 of disk unit 21, and physical sector 0 of disk unit 31 is stored in physical sector 0 of disk unit 31. Data D2 and its check code CRC2 are stored, respectively. The disk unit 41 of the parity disk device 4 has data disk devices 1, 2, 3, and 3.
The parity data of the data stored in the same physical sector and a check code of the parity data are stored. For example, in the physical sector 0 of the disk unit 41, the parity data CD012 of the data D0, D1, and D2 stored in the physical sector 0 of the disk devices 1, 2, and 3.
And the check code C of the parity data CD012
RC012 is stored.

A set of data stored in the same physical sector of the data disk devices 1, 2 and 3 is called logical data. For example, in the case of physical sector 0, data D0, D
The combination of 1 and D2 is one logical data. 1
When reading the entire logical data, all the disk devices 1, 2, 3, and 4 operate in parallel, the check by the check code is performed in each disk device 1, 2, 3, and 4, and the check by the parity data is performed. This is performed by the disk array control device 5. For example, when reading the entire logical data corresponding to physical sector 0, the disk device 1
In, the validity of the data D0 is verified from the data D0 read from the physical sector 0 and the check code CRC0. Similarly, data D1, D2, and CD012 read from physical sector 0 in disk devices 2, 3, and 4.
And its check code CRC1, CRC2, CRC01
2, the validity of the data D1, D2, CD012 is verified. If an error is detected in the data D0, D1, D2, CD012, if possible, the respective disk devices 1,
The data D0, D1, D2, and CD012 are corrected in 2, 3, and 4, and sent to the disk array controller 5. In the disk array controller 5, the read data D0,
After verifying the validity of the logical data from D1 and D2 and the parity data CD012, the logical data is sent to the host processor 6.

When reading only a part of the logical data, for example, the data D0 of the physical sector 0 of the disk device 1, the disk array control device 5 operates only the disk device 1. Also at this time, the disk device 1 verifies the validity of the read data D0 using the check code CRC0.

On the other hand, when updating the data, the disk array control device 5 operates the disk device 1 for the corresponding data.
Data is written to the disk devices 2 and 3 and the parity disk device 4. For example, when updating all the logical data corresponding to the physical sector 0 to the data D0 ', D1', D2 ', the disk array controller 5 sets the data D0'
, The check code CRC1 'from the data D1', the check code CRC2 'from the data D2', and the parity data CD01 from the data D0 ', D1', D2 '.
2 ′ and a check code CRC012 ′ of the parity data CD012 ′. Then, a request to write the data D0 'and the check code CRC0' to the physical sector 0 is issued to the disk device 1 via the disk bus 8-1, and the data D1 is written to the physical sector 0 to the disk device 2 via the disk bus 8-2. And a request to write a check code CRC1 'to the disk device 3 via the disk bus 8-3.
Request to write the data D2 'and the check code CRC2' to the disk device 4 through the disk bus 8-4 to the physical sector 0 in the parity data CD012 '.
And a request to write the check code CRC012 ′.

The disk controllers 12, 22, 32, and 42 in each of the disk devices 1, 2, 3, and 4 respond to the above-described write request to the disk units 11, 21, 31, and 41, respectively.
Is written to physical sector 0, but in the case of the present invention, prior to writing data to the disk unit,
Check the validity of the data. Then, if there is no error, writing to the disk unit is performed, but if there is an error, the writing operation is not performed.

The above is a case where the entire logical data is updated. However, when a part of the logical data is updated,
The disk array control device 5 includes disk devices 1, 2, 3, and 4 for corresponding data and a disk device 4 for parity.
Read and update the data recorded in, and perform rewriting. For example, when updating only the data D0 of the physical sector 0 of the disk device 1 to D0 ', the disk array control device 5 reads the data D0 from the disk device 1 and also reads the parity data CD012 from the disk device 4 and newly updates them. The parity data CD012 'is generated from the data D0' and the check code CRC012 'of the parity data CD012' is generated. Also, the check code CR from the new data D0 '
Generate C0 '. Then, a request to write the new data D0 'and the check code CRC0' to the physical sector 0 is issued to the disk device 1 through the disk bus 8-1.
The parity data CD012 'and the check code CRC01 are sent to the disk device 4 through the disk bus 8-4.
A request to write 2 'to physical sector 0 is issued.

The disk control units 12 and 42 in each of the disk devices 1 and 4 perform writing to the physical sector 0 of the disk units 11 and 41 in response to the above write request. Prior to writing to the disk, the validity of the data is checked, and after confirming that there is no error, writing to the disk unit is performed.

FIG. 2 is a block diagram showing a configuration example of the disk control unit 12. The disk control unit 12 of this example includes a data writing unit 121 that writes data received from the disk array control device 5 via the disk bus 8-1 to the disk unit 11 and a data read unit that reads data from the disk unit 11 and transmits the data via the disk bus 8-1. The data read unit 122 to be sent to the disk array control device 5 and the write request and read request received from the disk array control device 5 via the disk bus 8-1 are analyzed, and the data write unit 121 is written when a write request is made. The control unit 123 controls the data read unit 122.

The data writing unit 121 includes a buffer 1211 for temporarily storing write data (update data and its check code),
Write data checking means 1212 for checking the validity of the write data stored in the
And writing means 1213 for writing the write data held in the disk unit 1 to the disk unit 11. Then, the write data check unit 1212
1. a check code generator 12121 for generating a check code from the update data held in No. 1; a check code generated by the check code generator 12121;
The comparator 12 compares the check code held in the buffer 1211 with the check code and notifies the control unit 123 of the result.
122. Note that the check code generated by the check code generator 12121 is of the same type as the check code in the write data.

FIG. 3 is a flowchart showing a processing example of the control unit 123 when a data write request is issued from the disk array control device 5 to the disk device 1.
Hereinafter, the operation of the disk device 1 at the time of a write request will be described with reference to FIGS. For convenience of description, data D0 'and its check code CRC are stored in physical sector 0.
The case where 0 'is written is taken as an example.

The write data (D0 '+) transmitted from the disk array controller 5 through the disk bus 8-1
CRC0 ′) is stored in the buffer 1211 in the data writing unit 121 of the disk control unit 12, and a request to write the write data to the physical sector 0 is received by the control unit 123 via the disk bus 8-1. Write data check means 121 of data write section 121
2, the check code generator 12121 converts the data D0 ′ in the buffer 1211 into a check code CRC.
00 is output to one input terminal of the comparator 12122. The check code CRC0 ′ stored in the buffer 1211 is input to the other input terminal of the comparator 12122, and the comparator 12122 compares the two check codes and outputs a signal indicating match / mismatch to the control unit 123. I do.

When the request from the disk array controller 5 is a data write request, the control unit 123
Waiting for a comparison signal from 122, and when a signal indicating a comparison mismatch is output (NO in S1 of FIG. 3), an abnormal end is reported to the disk array controller 5 via the disk bus 8-1 ( S2), end the current process. On the other hand, when the signal indicating the comparison match is output (YES in S1), the writing destination is designated and the writing unit 1213 is specified.
Is started (S3). The writing unit 1213 reads the write data stored in the buffer 1211 and writes the data to the specified physical sector 0 of the disk unit 11. The control unit 123 is a writing unit 1213
Is completed, the normal completion is reported to the disk array control device 5 through the disk bus 8-1 (S4), and the current processing ends.

When a data read request is received from the disk array controller 5, the control unit 123 activates the data read unit 122, and the data read unit 122 reads data from the disk unit 11 and outputs data based on the check code. And sends the data to the disk array controller 5 via the disk bus 8-1. The processing at the time of reading this data is the same as in the related art.

The above is the description of the disk controller 1 of the disk device 1.
2 has been described, but the disk control units 22, 32, and 42 of the other disk devices 2, 3, and 4 also have the disk control unit 12
The configuration is the same as

FIG. 4 is a flowchart showing a processing example of the disk array controller 5 when receiving an update command from the host processor 6. The disk array controller 5 includes:
When an update instruction is received from the host processor 6 via the host bus 7, the instruction is analyzed (S11), and write data is generated (S12). For example, all the logical data corresponding to the physical sector 0 are converted into data D0 ', D1', D
2 ', the check code CRC0' is generated from the data D0 ', the check code CRC1' is generated from the data D1 ', the check code CRC2' is generated from the data D2 ', and the data D0', D1 ',
D2 ′, parity data CD012 ′ is generated, and the parity code CD012 ′ check code CR
C012 'is generated, and the data D0' and the check code CRC0 'are used as write data for the physical sector 0 of the disk device 1, and the data D1' and the check code CRC1 'are used as write data for the physical sector 0 of the disk device 2, Data D2 'and check code C
RC2 ′ is write data for the physical sector 0 of the disk device 3, and parity data CD012 ′ and the check code CRC012 ′ are write data for the physical sector 0 of the disk device 3. When only the data D0 of the physical sector 0 of the disk device 1 is updated to D0 ', the data D0 is read from the disk device 1 and the parity data CD012 is read from the disk device 4, and the parity data CD012 is read from these data and the new data D0'. Data C
D012 ', and the parity data CD
012 'check code CRC012' is generated.
Also, the check code CRC from the new data D0 '
Generate 0 '. Then, the new data D0 ′ and the check code CRC0 ′ are write data for the physical sector 0 of the disk device 1, and the parity data CD012 ′ and the check code CRC012 ′ are write data for the physical sector 0 of the disk device 4.

Next, the disk array control device 5 sends the generated write data and a write request designating the write destination to the corresponding disk devices 1, 2, 3, 4 to the disk buses 8-1, 8-2, 8-3. , 8-4 (S1).
3) Wait for a normal end report from all requested disk devices (S14). Then, when there is a report of the normal end from all the requested disk devices, the processing of FIG. 4 is ended. On the other hand, when waiting for reports of normal completion from all the requested disk devices, if there is a report of abnormal completion from any of the requested disk devices (YES in S15), the retry processing is performed. Proceed to (S16).
In this retry process, the same write request is sent again to all the requested disk devices or the disk device that has issued the abnormal end report. The retry process is repeated several times until it succeeds. If the retry is not successful even after retrying the predetermined number of times, for example, a notification to the effect that a failure has occurred in the disk device is given to the host processor 6. Upon receiving this notification, the host processor 6 outputs a failure message to, for example, a console (not shown).

[0034]

As described above, according to the present invention, before writing the data transferred from the disk array control device to the disk unit, the write data check means for checking the validity of the data is provided for each of them. Since the disk controller of the disk unit is provided, it is possible to detect immediately when data is corrupted due to a failure of the bus connecting the disk array controller and the individual disk units, and illegal data is written to the disk unit. This can be prevented from occurring.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a disk array device according to the present invention.

FIG. 2 is a block diagram illustrating a configuration example of a disk control unit.

FIG. 3 is a flowchart illustrating a processing example of a control unit in a disk control unit when a data write request is issued;

FIG. 4 is a flowchart illustrating a processing example of a disk array control device when an update command is received from a host processor.

FIG. 5 is an explanatory diagram of a conventional technique.

[Explanation of symbols]

 1, 2, 3, 4 ... disk device 11, 21, 31, 41 ... disk unit 12, 22, 32, 42 ... disk control unit 5 ... disk array controller 6 ... host processor 7 ... host bus 8-1, 8 -2,8-3,8-4 ... disk bus

Claims (5)

[Claims] 1. A disk array device comprising: a plurality of disk devices each including a disk control unit and a disk unit; and a disk array control device for controlling the plurality of disk devices. A disk array device comprising: a write data check unit for checking the validity of data before writing the data transferred from the disk array control device to the disk unit. 2. The write data checking means includes: means for generating a check code from data to be written to a disk unit; and a check code generated by the means and a check attached to the data transferred from the disk array controller. 2. The disk array device according to claim 1, further comprising means for comparing codes. 3. The disk control unit, when the validity of the data is confirmed by the data check unit, writes the data to the disk unit, reports a normal end to the disk array control device, and checks the data. 3. The disk array according to claim 1, wherein when the data abnormality is detected by the means, the disk array controller is notified of the abnormal termination without writing the data to the disk unit. apparatus. 4. The disk array device according to claim 3, wherein said check code is a CRC. 5. The disk according to claim 4, wherein said disk array control device performs a retry process when an abnormal end is reported from any of the disk devices that have requested data writing. Array device.